The goal of this proposal is to develop and validate practical methods which help physicians and dentists encourage their patients to stop smoking cigarettes and remain non-smokers. Parallel studies with physicians and dentists are proposed to determine the long-term effectiveness of practical smoking cessation interventions which they can use; and evaluate the adoption, generalizability and cost-effectiveness of implementing the "best" of those interventions with a new cohort of physicians and dentists. The research proposed incorporates and combines techniques used in our prior research in which: 1) explicit care protocols and feedback/reminder systems increased physicians' vigilance and management of a variety of medical problems; 2) nicotine-containing chewing gum produced significantly greater smoking cessation than did a placebo gum; and 3)\carbon monoxide (CO) from alveolar breath samples had equivalent specificity and greater sensitivity than did salivary thiocyanate for confirming self-reported smoking status. During State I, physicians and dentists will be randomly assigned to either, neither or both of the following: 1) a feedback/reminder system for following a practical protocol for helping their patients stop smoking; 2) nicotine gum freely available for their patients. Office exit interviews of patients at least every six months will determine changes in their smoking status confirmed by CO assessment and document their health professionals' behavior. The best Stage I cessation intervention will be used in Stage II with new cohorts of physicians and dentists randomly assigned to a control group or to implementing the "best" smoking cessation intervention either by relying primarily on the resources provided by our project staff (as in Stage I) or by relying primarily on the resources of their own regular practices. The major outcome is the percentage of a health professional's patients initially enrolled in the study who are not smoking at a particular unit of time (e.g., 6 months, 4 years later). Also assessed is the physicians'/dentists' adherence to the care protocols. Logistic regression is used to identify patient characteristics such as degree of cigarette dependence, expectations of quitting etc., most predictive of long-term cessation. Multiple regression is used to see if health professionals' characteristics such as smoking history, attitudes toward smoking intervention and expectation of success predict attainment of smoking cessation by their patients.